Comparative genomics suggests extensive antithrombotic gene expansion in Haemadipsa yanyuanensis Authors: Yiquan Lin, Fang Zhao, Sijia Fan, Dezhi Yang, Xiangrong Tong, Lizhou Tang, Dejun Kong, Gonghua Lin, Zichao Liu Published: 2025 ABSTRACT AbstractHaemadipsa yanyuanensis Liu & Song, 1977, is a terrestrial blood-feeding leech whose antithrombotic repertoire remains insufficiently characterized. We present its first chromosome-level genome assembly (165.32 Mb across nine chromosomes; BUSCO: 97.6%) using Nanopore long reads, Hi-C scaffolding, and RNA-seq data. We identified 193 putative antithrombotic genes in 15 putative families, representing a 2.2- to 2.7-fold increase in gene number but reduced family diversity compared with aquatic medicinal leeches, including Hirudo nipponia Whitman, 1886, Hirudinaria manillensis Lesson, 1842, and Hirudo medicinalis Linnaeus, 1758. Notably, the putative bdellin, LDTI, and LCI gene families exhibited marked lineage-specific expansions, ranging from 8.7- to 25-fold relative to those of aquatic leeches. In addition, the putative progranulin gene displayed a distinctive structural organization characterized by 122 cysteine residues and nine tandem repeats. Transcriptomic profiling revealed elevated expression of four expanded putative families, suggesting potential roles in terrestrial blood-feeding adaptation. This chromosome-scale genome provides a valuable resource for investigating anticoagulation mechanisms and may inform the development of next-generation antithrombotic therapeutics.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12864-025-12445-5.Keywords: Haemadipsa yanyuanensis, Terrestrial leech, Chromosome-scale genome, Antithrombotic genes, Gene family evolution Introduction Cardiovascular diseases (CVDs) represent one of the most pressing global public health challenges. According to the Global Burden of Disease study, the number of individuals affected by CVDs reached 523 million in 2019, with an associated mortality of 18.6 million cases [1]. Reports from the World Health Organization further indicate that CVD-related deaths accounted for 32% of all global mortality in 2022, with ischemic heart disease and cerebrovascular disorders contributing to approximately 85% of these fatalities [2]. Epidemiological projections suggest a concerning 90% rise in CVD prevalence between 2025 and 2050 [3]. Thrombosis, a central pathological mechanism in CVDs, involves activation of coagulation cascades and platelet-driven signaling pathways, ultimately leading to acute myocardial infarction and cerebral ischemia [4]. Given this substantial clinical burden, the development of novel antithrombotic therapies remains an urgent priority. Hematophagous annelids have emerged as valuable models for the discovery of thrombin-targeting therapeutics due to their evolutionarily refined antithrombotic systems. Hirudotherapy, an ancient medical practice, has been validated by modern pharmacological research [5, 6]. For instance, in 1989, recombinant hirudin variant 1 (HV1) from Hirudo medicinalis Linnaeus, 1758—expressed in Saccharomyces cerevisiae—was shown to retain the structural and biological properties of native hirudin in vitro and to exhibit potent antithrombotic activity in several in vivo thrombosis models [7]. Several recombinant hirudin analogs, including lepirudin, desirudin, and bivalirudin, have since been approved by the FDA and EMA as first-line treatments for heparin-induced thrombocytopenia [8, 9]. Although the antithrombotic mechanisms of aquatic medicinal leeches have been extensively investigated, the mechanisms in their terrestrial relatives—such as Haemadipsa yanyuanensis Liu & Song, 1977—remain poorly understood, limiting comparative analyses of blood-feeding adaptations across ecological lineages. Species within Haemadipsidae are known to possess a distinctive repertoire of hemostatic regulatory biomolecules. For example, the thrombin inhibitor haemadin, originally discovered in Haemadipsa sylvestris Blanchard, 1894, exhibits antithrombotic activity comparable to that of hirudin [10, 11]. Furthermore, pioneering work by Kvist et al. on Haemadipsa interrupta Moore, 1835, revealed substantial diversity in its antithrombotic components and provided insights into the evolutionary diversification of antithrombotic functions in medicinal leeches [12]. Subsequent studies identified additional haemadin-like genes in H. interrupta, further supporting the broad thrombin-inhibitory capacity of the haemadin family [13]. However, the molecular basis of anticoagulation in H. yanyuanensis remains largely unclear. Although two antithrombotic proteins have been detected in this species [14], their complete amino acid sequences have not yet been elucidated. Recent genomic investigations have characterized the putative antithrombotic gene repertoires of several aquatic medicinal leeches, including H. medicinalis [15], Hirudinaria manillensis Lesson, 1842 [16], Whitmania pigra Whitman, 1886 [17], Hirudo nipponia Whitman, 1884 [18], and Hirudinaria bpling Wahlberg, 1855 [19]. In contrast, terrestrial Haemadipsidae species lack comparable genome-scale resources, hindering genomic and evolutionary comparisons across ecological contexts. Notably, a genome assembly for the cave-dwelling leech Sinospelaeobdella cavatuses Yang, Mo & Wang, 2009 (Haemadipsidae) was recently published [20]; however, the study did not explore its functional genomics. In the present study, we employed the BRAKER-plus integrative annotation framework to generate the first high-quality genome assembly for H. yanyuanensis and to predict its putative antithrombotic genes, thereby filling this critical knowledge gap. Our comparative genomic analyses identified four major modules of evolutionarily conserved putative antithrombotic genes: (1) coagulation cascade inhibitors, (2) platelet activation antagonists, (3) fibrinolytic enhancers, and (4) tissue penetration facilitators. Together, these findings shed new light on the evolutionary adaptation of terrestrial annelids and highlight their potential as promising sources of novel antithrombotic and thrombolytic agents. Materials and methods DNA and RNA sequencing Specimens of H. yanyuanensis were collected from Jiaozi Snow Mountain, Luquan County, Yunnan Province, China (26.023°N, 102.872°E). Following a seven-day starvation period and subsequent removal of the digestive tract, paired-end Illumina libraries and ultra-long Nanopore libraries were prepared according to established protocols [18, 21]. This sequencing workflow generated ultralong Oxford Nanopore Technologies (ONT) reads for de novo genome assembly, alongside strand-specific RNA-seq libraries for transcriptomic analysis. Genomic and transcriptomic sequencing was performed following previously described methods, and high-quality data were obtained using 150 bp paired-end sequencing on the Illumina HiSeq 2000 platform (Illumina Inc., San Diego, CA, USA). Genome assembly and gene prediction The genome of H. yanyuanensis was assembled following previously described protocols [18], and putative antithrombotic genes were subsequently identified. Briefly, preliminary nuclear and mitochondrial genome assemblies were generated using NextDenovo v2.5.0 in combination with a comprehensive suite of bioinformatic tools. Chromosome-level scaffolding was accomplished using Hi-C data, and topological interaction maps were visualized using standard methods [22, 23]. Genome completeness was evaluated using BUSCO v4.1.4 against the eukaryota_odb10 database, whereas assembly accuracy was assessed using Merqury v1.3. A de novo repeat library was constructed using RepeatModeler v2.0.3 and the RepBase v20181026 database [24], and repetitive elements were masked with RepeatMasker v4.1.2-p1. Gene annotation was performed using the BRAKER-plus pipeline [25]. RNA-seq reads were first aligned to the genome assembly, followed by de novo transcript reconstruction using Trinity v2.9.0. Coding sequences (CDSs) were predicted using GeneMarkS-T v5.1. A reference library of putative antithrombotic genes was constructed through literature review and mining of public repositories such as GenBank and UniProt. CDSs from the RNA-seq and BRAKER predictions were compared against this library using BLASTP to identify putative homologs, and filtered candidate genes were mapped to genomic loci using Exonerate v2.2.0. After manual refinement, BRAKER predictions were merged with the GFF annotation file using AGAT v1.2.0, yielding a comprehensive annotation set for putative antithrombotic genes. Signal peptide regions of candidate proteins were predicted using SignalP v6.0 [26] (accessed April 13, 2025). To investigate phylogenetic relationships among gene family members, multiple sequence alignment was performed in MEGA v11.0.13, followed by maximum likelihood tree construction in IQ-TREE v1.6.12 with 1,000 bootstrap replicates. Genetic variation analysis of antithrombotic genes Putative antithrombotic protein-coding genes identified in this study were combined with corresponding gene sequences from H. medicinalis, H. manillensis, S. cavatuses, and H. nipponia, together with prototype sequences. Nucleotide sequences were translated into amino acid sequences using the “Translated Protein Sequences” function in MEGA with the default standard genetic code. The resulting protein sequences, including prototype sequences, were aligned using the ClustalW algorithm in MEGA under default parameters. A maximum likelihood phylogenetic tree was subsequently constructed in IQ-TREE to examine the evolutionary relationships among members of the HMEI gene family. To investigate gene expression variation, RNA-seq data from H. yanyuanensis were integrated with reference CDS sequences. Gene expression levels were quantified using Salmon v1.0.0, and transcripts per million (TPM) values were calculated. TPM values from H. yanyuanensis were then integrated with published expression datasets for the blood-feeding leeches H. nipponia and H. manillensis to enable cross-species comparative analyses. Because corresponding RNA-seq data are not available for H. medicinalis, this species was excluded from the comparative expression analyses. Results Gene sequencing and assembly In this study, we generated 15.18 Gb of high-quality Oxford Nanopore Technologies (ONT) long-read data with an average read length of 11.65 kb. De novo assembly produced 37 contigs totaling 169.56 Mb, with a contig N50 of 7.92 Mb. Hi-C sequencing on the Illumina HiSeq platform generated 43.48 Gb of data, which enabled chromosome-scale scaffolding. The contigs were assembled into 18 scaffolds totaling 165.30 Mb, with an N50 of 18.42 Mb. Scaffold length distribution analysis (Fig. 1) indicated that the nine longest scaffolds represent nine chromosomes, together comprising 164.47 Mb (99.5%) of the assembly. The remaining nine scaffolds are smaller than 0.3 Mb. Additionally, 20.41 Gb of Illumina paired-end reads were used for mitochondrial genome assembly with GetOrganelle, resulting in a 17,851 bp circular mitochondrial genome. Altogether, the final H. yanyuanensis genome assembly spans 165.32 Mb, including nine chromosomes, one mitochondrial genome, and nine unplaced scaffolds. Scaffold and chromosome lengths of Haemadipsa yanyuanensis Liu& Song, 1977, genome To assess the completeness of the genome assembly, BUSCO (eukaryota_odb10) identified 97.6% of conserved orthologous genes, with 2.4% of genes missing, of which 91.8% were complete single-copy genes. Merqury analysis produced a quality score of 39.03, indicating high base-level accuracy. Repeat annotations indicated that 28.49% of the genome consists of putative repetitive elements, including retrotransposons and DNA transposons (Table 1). No small RNAs, satellite sequences, or low-complexity regions were reliably detected in this assembly. Table 1Different repeat sequence type percentages in genomes of Haemadipsa yanyuanensis Liu & Song, 1977 (Hyan), Hirudinaria manillensis Lesson, 1842 (Hman), and Hirudo nipponia Whitman, 1886 (Hnip)ItemHyanHmanHnipRetroelements9.516.599.89DNA transposons1.981.768.40Rolling-circles1.060.200.18Unclassified13.178.4110.71Small RNA0.000.000.00Satellites0.000.200.02Simple repeats2.771.800.58Low complexity0.000.000.00Total28.4918.9629.77 Different repeat sequence type percentages in genomes of Haemadipsa yanyuanensis Liu & Song, 1977 (Hyan), Hirudinaria manillensis Lesson, 1842 (Hman), and Hirudo nipponia Whitman, 1886 (Hnip) Gene prediction and annotation A systematic analysis of anticoagulant biomolecules in leech saliva revealed five putative antithrombotic gene clusters and 23 putative antithrombotic proteins. Using the BRAKER-plus pipeline, we performed a comprehensive structural genome annotation for H. yanyuanensis, resulting in 26,509 predicted genes. BUSCO (eukaryota_odb10) analysis indicated a high level of completeness, with 95.3% of genes identified as complete, 3.1% as fragmented, and only 1.6% missing, suggesting that the CDS annotation is largely comprehensive. In total, 15 putative gene families and 193 putative antithrombotic gene loci were identified, including three predicted pseudogenes. Comparative BLASTP analysis using CDS sequences from S. cavatuses identified 109 putative antithrombotic genes in this species. Due to high sequence similarity among Haemadipsa species, proteins such as hirustasin, hirustasin-like, guamerin, piguamerin, bdellastasin, and poecistasin could not be reliably distinguished and were therefore merged into the leech inhibitor superfamily. Ultimately, 18 putative antithrombotic-related gene/protein families were identified. These gene families include eight predicted coagulation inhibitors (haemadin [10], progranulin [27, 28], antistasin [29], lefaxin [30], hirustasin [31], eglin [32], bdellin B-3 [33], and LDTI [34, 35]). Three fibrinolytic enhancers (destabilase [36], GGT [37], and LCI [38]) and two platelet aggregation inhibitors (apyrase [39] and lumbrokinase [40]) were also detected, along with one tissue-permeability enhancer (hyaluronidase [41]). Notably, a HMEI gene family comprising 53 members was identified, which may function as a protease inhibitor with potential anti-inflammatory activity [42, 43]. In contrast, no confident gene members of the therostasin or saratin families were found in H. yanyuanensis, suggesting a possible lineage-specific gene loss event. Furthermore, decorsin family members were detected exclusively in H. nipponia (Table 2). Table 2Number of functional genes of H. yanyuanensis (Hyan), Sinospelaeobdella cavatuses Yang, Mo & Wang, 2009 (Scav), H. nipponia (Hnip), H. manillensis (Hman) and Hirudo medicinalis Linnaeus, 1758 (Hmed)Gene familyHyanScavHnipHmanHmedFunctionhaemadin/hirudin30353coagulation inhibitorprogranulin11111coagulation inhibitorantistasin31222coagulation inhibitorlefaxin22333coagulation inhibitortherostasin00111coagulation inhibitorhirustasin275191813coagulation inhibitoreglin137545coagulation inhibitorbdellin269312coagulation inhibitorLDTI168111coagulation inhibitordecorsin00300platelet aggregation inhibitorsaratin00221platelet aggregation inhibitorapyrase22655platelet aggregation inhibitorlumbrokinase735332platelet aggregation inhibitordestabilase94438fibrinolysis enhancerGGT21412fibrinolysis enhancerLCI251111fibrinolysis enhancerhyaluronidase44434tissue penetration enhancerHMEI5329211820inflammation inhibitortotal193109867274—— Number of functional genes of H. yanyuanensis (Hyan), Sinospelaeobdella cavatuses Yang, Mo & Wang, 2009 (Scav), H. nipponia (Hnip), H. manillensis (Hman) and Hirudo medicinalis Linnaeus, 1758 (Hmed) Protein and genetic variation analysis Phylogenetic analyses resolved blood-feeding leeches and mountain leeches into two well-supported clades (Fig. 3C). In comparative analyses of the putative antithrombotic gene families in H. yanyuanensis, S. cavatuses, and three reference medicinal leeches (H. nipponia, H. manillensis, and H. medicinalis), we observed extensive genetic variation among members of each putative family (Figs. S1–S18). These comparisons revealed four potential evolutionary patterns, consisting of two expansion modes—copy-number amplification and coding-sequence elongation—and two contraction modes involving reductions in these features. Comparative transcriptomic expression profiles of putative antithrombotic genes in leeches. A–F Median TPM line plots of six putative antithrombotic gene families (lumbrokinase, eglin, LCI, bdellin, HMEI, and hirustasin) in H. nipponia (Hnip), H. manillensis (Hman), and H. yanyuanensis (Hyan). G Hierarchical clustering heatmap of putative antithrombotic gene expression based on RNA-seq datasets from four leech species (blue indicates higher similarity). The datasets include H. nipponia (Hnip1–Hnip3, SRR26541742–SRR26541740), H. manillensis (Hman1–Hman3, SRR26541746, SRR26541753, SRR26541752), H. yanyuanensis (Hyan1–Hyan3, CRR1847694–CRR1847699), and S. cavatuses (Scav1, SRR28904419) The first expansion pattern involved nine putative gene families, including antistasin, hirustasin, eglin, bdellin, LDTI, lumbrokinase, destabilase, LCI, and HMEI. Among these families, eglin (13 members), lumbrokinase (7), and HMEI (53) exhibited approximately twice as many members as observed in the three reference medicinal leeches. Notably, the HMEI family contained one predicted pseudogene (HMEI_Hyan52) and another gene (HMEI_Hyan11) harboring multiple nonsynonymous substitutions and a premature stop codon in the C-terminal region (Fig. S1). The HMEI family accounted for 27.5% of all identified putative antithrombotic-related genes (n = 193; Table 2). Substantial expansions were also observed in the LDTI, bdellin, and LCI families. For example, 26 bdellin genes were annotated in H. yanyuanensis, including a predicted pseudogene (bdellin_Hyan26) containing a premature stop codon in the signal peptide region (FiG. S2), whereas the three reference medicinal leeches contained only 3, 1, and 2 bdellin genes, representing an expansion of 8.7- to 26-fold. Similarly, the LDTI (Fig. S3) and LCI (Fig. S4) families contained 16 and 25 members, respectively, compared with a single-copy ortholog in each medicinal leech species, corresponding to 16-fold and 25-fold increases. Other families, such as destabilase (Fig. S5), exhibited more moderate expansion and also included a predicted pseudogene (destabilase_Hyan09). In sharp contrast, the putative lefaxin and apyrase families underwent contraction in H. yanyuanensis. The apyrase family contained only two members (Fig. S6), whereas the three reference medicinal leeches possessed six, five, and five members, respectively. Similarly, the lefaxin family in H. yanyuanensis retained only two members (Fig. S7), slightly fewer than the three present in each of the reference medicinal leeches. Comparable patterns of expansion and contraction were also observed in H. interrupta [13] and S. cavatuses. In H. interrupta, the putative haemadin family exhibited a marked expansion, with copy numbers reaching 12 (Fig. 2D). In S. cavatuses, the overall pattern of putative family expansion and contraction was broadly similar to that of H. yanyuanensis. Notably, the putative lumbrokinase family expanded dramatically to 35 members, far exceeding the levels reported in any other related leech species to date (Table 2). However, unlike H. yanyuanensis, S. cavatuses appeared to lack detectable haemadin genes, showed contraction in the hirustasin and antistasin families, and exhibited no expansion in the LCI family. Comparative analysis of different leech species and their relationship between hirudin and haemadin protein families. A Multiple sequence alignment of hirudin and haemadin proteins from four leech species (Hirudo nipponia Whitman, 1886 (Hnip), Hirudo medicinalis Linnaeus, 1758 (Hmed), Hirudinaria manillensis Lesson, 1842 (Hman), and H. yanyuanensis (Hyan)), with the prototype hirudin (GenBank accession number: ALA22933.1) listed. B Sequence length of haemadin_Hyan3 and the internal tandem repeat sequence of the haemadin protein. C Phylogenetic relationships of selected leech species (CO1). D Phylogenetic tree of hirudin and haemadin proteins. Sequences Hint_v1–Hint_v12 correspond to haemadins identified from Haemadipsa interrupta Moore, 1835. The red boxes highlight the signal peptide regions; red triangles indicate conserved cysteine residues; black asterisks denote stop codons; purple represents hirudin; blue represents referenced haemadins; and red represents haemadins from H. yanyuanensis The second expansion mode was characterized by substantial elongation of putative protein-coding sequences. In several cases, the predicted protein products were nearly twice as long as those of their orthologs in the reference medicinal leeches. This pattern was observed in four putative gene families: LDTI (Fig. S3), haemadin (Fig. 2 A), progranulin (FiG. S8), and antistasin (Fig. S9), yielding six putative antithrombotic polypeptides: LDTI_Hyan1 through LDTI_Hyan3, haemadin_Hyan3, progranulin_Hyan, and antistasin_Hyan1. For example, haemadin_Hyan3 contained three internal tandem repeats, each consisting of six cysteine residues (Fig. 2B), and antistasin_Hyan1 included five repeats, each with ten conserved cysteines (Fig. S10). Phylogenetic analysis further suggested a close evolutionary relationship between the haemadin_Hyan3 protein and the Hint_v6 and Hint_v7 proteins (Fig. 2D). The progranulin gene was conserved as a single-copy locus across all five species, but the number of cysteine residues varied substantially: H. nipponia possessed 69, H. manillensis 66, and H. medicinalis 72, whereas H. yanyuanensis and S. cavatuses contained as many as 122 and 137, respectively (Figs. S8, S11). Moreover, the progranulin protein sequences of H. yanyuanensis and S. cavatuses comprised nine and eleven conserved tandem repeat units, respectively, each consisting of twelve cysteine residues, whereas H. nipponia and H. medicinalis harbored only five and six such repeat units [16, 18] (Fig. S12). In contrast to this elongation pattern, the bdellin gene family exhibited contraction of protein-coding sequences. Despite the marked increase in gene copy number, 22 bdellin genes encoded truncated proteins that were approximately half the length of their orthologs in the reference medicinal leeches (Fig. S2), suggesting possible functional divergence or degeneration. Gene expression analysis By integrating publicly available datasets with the transcriptomic profiles of H. yanyuanensis, we detected notable interspecific variation in putative antithrombotic gene expression. Transcriptomic analyses combined with putative gene family expansion patterns suggested that the six families showing relatively high expression levels in H. yanyuanensis—lumbrokinase, LCI, bdellin, HMEI, hirustasin, and eglin—were all members of expanded gene families. Among these, the lumbrokinase, LCI, bdellin, and HMEI families exhibited markedly higher expression in H. yanyuanensis than in H. nipponia and H. manillensis (Fig. 3A and D). In contrast, the eglin family displayed an intermediate expression pattern, with expression levels higher than those of H. nipponia but lower than those of H. manillensis (Fig. 3E). Conversely, the hirustasin family exhibited the opposite trend, with expression levels exceeding those of H. manillensis but remaining lower than those of H. nipponia (Fig. 3F). To further assess whether the expression differences of the six putative antithrombotic-related gene families were statistically significant among the three leech species, we performed Mann–Whitney U tests. Each gene family was represented by three samples per species, and the results identified nine significant pairwise comparisons (Table 3). For the lumbrokinase family, a significant difference was observed between H. manillensis and H. yanyuanensis. In contrast, the LCI and bdellin families exhibited consistent significant differences across all pairwise comparisons. For the HMEI family, a significant difference was detected exclusively between H. nipponia and H. yanyuanensis, whereas no significant differences were identified among species for the eglin family. Conversely, the hirustasin family showed a significant difference only between H. nipponia and H. manillensis. Table 3Mann–Whitney U test results for the expression levels of six putative antithrombotic gene families among H. nipponia (Hnip), H. manillensis (Hman), and H. yanyuanensis (Hyan). Significant results (P < 0.050) are shown in bold. All P values are reported to three decimal placesGene familyComparison groupZ P lumbrokinaseHnip vs. Hman−1.1590.246Hnip vs. Hyan−1.0910.275Hman vs. Hyan −1.993 0.046 LCIHnip vs. Hman −1.964 0.049 Hnip vs. Hyan −1.964 0.049 Hman vs. Hyan −1.964 0.049 bdellinHnip vs. Hman −1.964 0.049 Hnip vs. Hyan −1.964 0.049 Hman vs. Hyan −1.964 0.049 HMEIHnip vs. Hman−0.6550.512Hnip vs. Hyan −1.964 0.049 Hman vs. Hyan−1.0910.275eglinHnip vs. Hman−1.5280.126Hnip vs. Hyan−1.5280.126Hman vs. Hyan−0.6550.512hirustasinHnip vs. Hman −1.964 0.049 Hnip vs. Hyan−1.0910.512Hman vs. Hyan−1.0910.275 Mann–Whitney U test results for the expression levels of six putative antithrombotic gene families among H. nipponia (Hnip), H. manillensis (Hman), and H. yanyuanensis (Hyan). Significant results (P < 0.050) are shown in bold. All P values are reported to three decimal places In the clustering heatmap, the bdellin gene family displayed markedly elevated expression levels, substantially higher than those reported for any other putative antithrombotic-related gene family previously characterized in medicinal leeches. In contrast, the antistasin and lefaxin families showed lower expression in H. yanyuanensis relative to H. nipponia and H. manillensis. Hierarchical clustering further indicated species-specific transcriptomic patterns. Consistent with their phylogenetic relationships, H. yanyuanensis, H. nipponia, and S. cavatuses formed a closely related clade. Moreover, the expanded lumbrokinase, eglin, and bdellin families in S. cavatuses also exhibited elevated TPM expression (Fig. 3G). With the exception of the outlier sample Hnip2—likely reflecting biological or technical variation—the overall clustering pattern suggested that eight putative antithrombotic gene families (bdellin, hirustasin, eglin, lefaxin, destabilase, LCI, HMEI, and LDTI) consistently maintained relatively high TPM expression across the four species. Discussion Genome assembly In this study, we employed whole-genome sequencing to investigate the evolutionary divergence between H. yanyuanensis and three reference medicinal leeches (H. medicinalis, H. manillensis, and H. nipponia) as well as S. cavatuses, thereby uncovering lineage-associated genomic differences. By combining Oxford Nanopore Technologies (ONT) long reads with Illumina Hi-C, short-read survey sequencing, and RNA-seq data, we generated a high-quality, chromosome-level assembly of 165.31 Mb resolved into nine chromosomes, together with a nearly complete mitochondrial genome. The assembly was strongly supported by a BUSCO completeness of 95.9% and a Merqury quality score of 39.9. Gene expansion–contraction model Comparative analysis revealed that H. yanyuanensis harbors 193 putative antithrombotic genes—approximately 2.2 to 2.7 times more than those identified in the medicinal leeches (H. nipponia, 86; H. medicinalis, 74; H. manillensis, 72)—yet these genes are distributed among only 15 families, compared to 17–18 in the reference medicinal leeches [16, 18]. Moreover, 109 putative antithrombotic genes mined from the CDS sequences of S. cavatuses also exceeded those detected in the medicinal leeches. This pattern suggests lineage-associated remodeling through differential expansion and contraction of gene families. Notably, the bdellin, LDTI, and LCI families showed substantial expansions, with up to 26-fold more members than in the reference species, whereas the apyrase and lefaxin families exhibited marked reductions. Several expanded families also included pseudogenes or truncated sequences, indicating potential functional divergence. A similar expansion–contraction trend was observed when comparing H. yanyuanensis with non-medicinal leeches [21]. In addition to changes in copy number, several proteins (such as haemadin_Hyan3 and progranulin_Hyan) exhibited pronounced sequence elongation caused by internal tandem repeats and elevated cysteine content. For example, progranulin_Hyan contains 122 cysteine residues distributed across nine conserved repeat units, nearly twice the number found in its homologs, suggesting a more complex structural organization. Together, these changes indicate that H. yanyuanensis may have adopted a unique combination of gene family expansion, contraction, and domain proliferation. Through gene duplication, domain extension, and potential neofunctionalization [44], these processes may contribute to its enlarged repertoire of antithrombotic factors and could be associated with lineage-specific ecological pressures. To our knowledge, such a pattern has not been previously reported among terrestrial leeches. Environment-associated expansion patterns were also observed in H. interrupta, which contains up to 12 haemadin genes—exceeding the numbers reported in other leech species. One possible explanation is that H. yanyuanensis and H. interrupta inhabit environments with markedly different temperature ranges, which may impose distinct selective pressures on haemadin gene retention. However, this hypothesis remains speculative and requires further experimental validation. A comparable pattern was found in S. cavatuses, in which no haemadin gene was identified from the CDS data. Given that S. cavatuses belongs to the genus Sinospelaeobdella (Haemadipsidae) and represents a specialized terrestrial lineage, additional data will be necessary to clarify the evolutionary basis of this pattern. Differences in gene family expression levels Due to the absence of RNA-seq data for H. medicinalis, the comparative expression analysis was limited to H. yanyuanensis, H. nipponia, and H. manillensis. In H. yanyuanensis, four of the six expanded putative gene families (eglin, bdellin, HMEI, and LCI) showed higher expression levels than those observed in the other two leech species. Although putative LDTI and hirustasin did not completely follow this trend, the overall results suggest that gene family expansion may have contributed to the elevated expression levels. It should be noted that half of the 18 pairwise comparisons among the six expanded putative gene families did not show significant differences, likely due to low statistical power associated with the limited number of biological replicates per species. Future studies will require larger sample sizes—ideally ten or more biological replicates—along with further experimental validation. Taken together, the increased expression of several antithrombotic putative gene families may also be associated with the terrestrial lifestyle of the genus Haemadipsa. Biomedical potential and conservation considerations of Haemadipsa species Recent research on terrestrial leeches has largely focused on the use of invertebrate-derived DNA (iDNA) for vertebrate biodiversity monitoring [45, 46]. At the same time, pharmacological studies have increasingly emphasized their antithrombotic potential, with compounds such as hirudin and the newly identified sylvestin demonstrating potent biological activity [11, 47]. Functional validation of decorin and other haemadin-like molecules in H. interrupta further supports their clinical relevance [13]. Collectively, these findings position members of the genus Haemadipsa, including H. yanyuanensis, as promising sources for the discovery of novel antithrombotic agents. Although habitat degradation has contributed to population declines in aquatic leeches [48–51], terrestrial Haemadipsa species—inhabiting tropical and subtropical forests—have remained relatively stable due to their broad geographic distribution. New species continue to be identified within this group [52]. The ecological resilience of the genus, combined with the lineage-specific expansion of key antithrombotic gene families revealed in this study, underscores both its biomedical potential and its value for long-term drug discovery efforts. Conclusions This study presents the first chromosome-level genome assembly of the terrestrial leech Haemadipsa yanyuanensis, revealing a distinctive evolutionary trajectory of putative antithrombotic gene adaptation. Although H. yanyuanensis harbors fewer antithrombotic gene families (15) than aquatic medicinal leeches (17–18), it contains a substantially larger number of putative antithrombotic genes (193). Several families (such as bdellin, LDTI, and LCI) display extreme copy-number expansion, reaching up to 26-fold increases relative to reference species. Furthermore, elongation of coding sequences and the presence of extensive tandem-repeat structures in families such as progranulin and haemadin indicate additional functional diversification potentially shaped by terrestrial selective pressures. Overall, these findings support the possibility that coordinated gene family contraction and expansion contribute to lineage-specific adaptive trajectories within Haemadipsa, potentially linked to ecological conditions associated with terrestrial blood feeding. By narrowing an existing genomic gap for terrestrial leeches, this work provides a foundation for future evolutionary analyses and, with appropriate functional validation, may offer preliminary molecular leads relevant to antithrombotic drug discovery. Further experimental and ecological studies will be essential to fully clarify these hypotheses. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4. Supplementary Material 5. Supplementary Material 6. Authors’ contributions Conceptualization, Z. L., G. L., D. K., and Y. L.; Methodology, Y. L. and G. L.; Software, Y. L. and G. L.; Validation, Formal Analysis, Investigation, Resources, Data Curation, and Visualization, Z. L., G. L., Y. L., F. Z., S. F., D. Y., X. T., and L. T.; Writing and Editing, Z. L., G. L., and Y. L.; Supervision, Project Administration, and Funding Acquisition, Z. L., G. L., and Y. L. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the National Natural Science Foundation of China (82260742 and 32260132), the Yunnan Provincial University Serving Key Industry Science and Technology Special Project (FWCY-ZD2024009), the Joint Special Project for Basic Research of Local Universities in Yunnan Province (202301BA070001-105 and 202101BA070001-164), the 2023 Frontier Research Team Program of Kunming University, the Key Laboratory of Jiangxi Province for Biological Invasion and Biosecurity (2023SSY02111), National Key R&D Program of China (2023YFF1305000), the Yunnan Ten Thousand Talents Program for Young Top Notch Talents (YNWR-QNBJ-2020-122), Special Project of Science-Technology Talent and Platform in Yunnan Province (202105AC160042), and Yunnan International Joint Laboratory with South and Southeast Asia for the Integrated Development of Animal-Derived Anti-Thrombosis Chinese Medicine (202503AP140025). Data availability The raw genomic data generated in this study have been deposited at Genome Sequence Archive (GSA) portal of the China National Center for Bioinformation **(https://ngdc.cncb.ac.cn/gsa)** with accession number of CRA025628. This repository is maintained by the Chinese Academy of Sciences, following international standards for biological data sharing. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Dejun Kong, Email: kongdj@ynu.edu.cn. Gonghua Lin, Email: lingonghua@163.com. Zichao Liu, Email: abclzc@aliyun.com. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4. Supplementary Material 5. Supplementary Material 6. Data Availability Statement The raw genomic data generated in this study have been deposited at Genome Sequence Archive (GSA) portal of the China National Center for Bioinformation **(https://ngdc.cncb.ac.cn/gsa)** with accession number of CRA025628. This repository is maintained by the Chinese Academy of Sciences, following international standards for biological data sharing.